Automobile fault diagnosis and prediction maintenance system and method based on fault tree analysis method

Abstract

The invention relates to the technical field of automobile fault diagnosis and maintenance, and particularly discloses an automobile fault diagnosis and prediction maintenance system and method based on a fault tree analysis method. The system comprises a fault tree construction module which constructs a fault tree of a to-be-analyzed fault phenomenon based on a top event, all intermediate events and all basic events of the to-be-analyzed fault phenomenon; the minimum cut set determination module determines all minimum necessary cut sets of the top event based on the fault tree of the fault phenomenon to be analyzed; the hidden danger prediction and evaluation module calculates occurrence probability time sequence data of all middle events and top events based on the loss trend data of the automobile parts, and analyzes a relative hidden danger level of a fault tree of a fault phenomenon to be analyzed in a prediction period in combination with all minimum necessary cut sets of the top events; a maintenance strategy generation module generates a maintenance strategy based on the fault tree of the fault phenomenon to be analyzed in the relative hidden danger level of the prediction period; automobile maintenance resources and opportunities are reasonably arranged, and faults are effectively prevented.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile fault diagnosis and maintenance, and in particular to an automobile fault diagnosis and predictive maintenance system and method based on a fault tree analysis method. Background Art

[0002] Currently, with the rapid development of the automotive industry, the number of vehicles on the road continues to rise, and vehicle reliability and safety are receiving increasing attention. As complex mechatronic products, vehicles feature numerous components and complex system structures. Consequently, various faults are inevitable over the course of long-term use. Once a fault occurs, it not only impacts the vehicle's normal operation but can also pose a threat to the safety of the driver and passengers. Therefore, accurate and efficient automotive fault diagnosis and predictive maintenance technologies are crucial. Fault diagnosis can quickly locate the cause of a fault, enabling maintenance personnel to promptly repair the vehicle, reducing downtime and repair costs. Predictive maintenance, on the other hand, uses real-time monitoring and analysis of the vehicle's operating status to predict potential faults in advance, shifting maintenance from reactive repairs to proactive prevention. This effectively reduces the probability of failures, extends the vehicle's service life, and improves its overall performance and safety.

[0003] However, existing automotive fault diagnosis and predictive maintenance technologies suffer from numerous problems. The lack of a systematic approach to aligning the logical relationships between fault symptoms and various levels of events results in in-depth and inaccurate analysis of fault causes. Furthermore, the failure to consider wear trends of automotive components and the probabilistic derivation of all events related to the fault symptoms during the potential hazard prediction and assessment phase results in inaccurate and unreliable predictions. This ultimately makes it difficult to formulate scientifically sound maintenance strategies and effectively predict and prevent vehicle failures.

[0004] Therefore, the present invention proposes an automobile fault diagnosis and predictive maintenance system and method based on the fault tree analysis method. Summary of the Invention

[0005] The application provides an automobile fault diagnosis and predictive maintenance system and method based on a fault tree analysis method, the system constructs a fault tree based on top events, intermediate events and basic events by using a fault tree construction module, provides a clear logical framework for the entire fault analysis, intuitively presents the hierarchical relationship between fault causes, comprehensively combs the logical relationship between fault phenomena and events at each level, and facilitates comprehensive and systematic understanding of the fault structure. A minimum cut set determination module accurately locates the key factor combination causing the fault by determining the minimum necessary cut set of the top event, and focuses on the core points for subsequent analysis. Furthermore, a hidden danger prediction and evaluation module calculates event occurrence probability time series data in combination with automobile component wear trend data, and analyzes the relative hidden danger level in combination with the minimum necessary cut set, realizes dynamic evaluation and prediction of the fault hidden danger, perceives potential risks in advance, and greatly improves the accuracy and reliability of the prediction results. A maintenance strategy generation module generates a scientific and reasonable maintenance strategy according to the relative hidden danger level, makes the maintenance work more targeted and forward-looking, reasonably arranges maintenance resources and time, effectively prevents faults, guarantees the safety of automobile operation, reduces maintenance costs, and improves the overall reliability and service life of the automobile.

[0006] The application provides an automobile fault diagnosis and predictive maintenance system based on a fault tree analysis method, comprising:

[0007] A fault tree construction module is used to construct a fault tree of a fault phenomenon to be analyzed based on a top event and all intermediate events and all basic events of the fault phenomenon to be analyzed;

[0008] A minimum cut set determination module is used to determine all minimum necessary cut sets of the top event based on the fault tree of the fault phenomenon to be analyzed;

[0009] A hidden danger prediction and evaluation module is used to calculate occurrence probability time series data of all intermediate events and the top event based on wear trend data of automobile components, and analyze the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in a prediction period in combination with all minimum necessary cut sets of the top event;

[0010] A maintenance strategy generation module is used to generate a maintenance strategy based on the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period.

[0011] Preferably, the fault tree construction module comprises:

[0012] An event determination sub-module is used to determine the top event and all intermediate events and all basic events based on automobile operation related data of the fault phenomenon to be analyzed;

[0013] A fault tree building sub-module is used to build a tree structure of the top event and all intermediate events and all basic events of the fault tree of the fault phenomenon to be analyzed based on the logical relationship between all events, and obtain the fault tree of the fault phenomenon to be analyzed.

[0014] Preferably, the minimum cut set determination module comprises:

[0015] The top event expression construction submodule is configured to represent all basic events in the fault tree with Boolean variables and perform a step-by-step upward logical deduction in the fault tree to obtain a Boolean expression of the top event.

[0016] The expression simplification and expansion submodule is configured to expand the Boolean expression of the top event based on preset simplification rules to obtain a plurality of basic event product terms, and treat each basic event product term as a single minimum necessary cut set of the top event.

[0017] Preferably, the hazard prediction and evaluation module comprises:

[0018] The fault probability evaluation submodule is configured to determine occurrence probability time series data of all basic events in a prediction period based on the wear trend data of the automobile components.

[0019] The event probability deduction submodule is configured to perform an upward deduction on the occurrence probability time series data of all basic events in the prediction period based on the fault tree to obtain occurrence probability time series data of all intermediate events and the top event.

[0020] The fault synchronicity value evaluation submodule is configured to analyze fault synchronicity values of all minimum necessary cut sets of the top event based on the occurrence probability time series data of all basic events.

[0021] The guidance performance evaluation submodule is configured to assign values to each level in the fault tree based on the fault synchronicity values of all minimum necessary cut sets of the top event to obtain decision guidance comprehensive performance values of each level in the fault tree.

[0022] The hazard prediction and evaluation submodule is configured to determine a relative hazard level of the fault tree of the fault phenomenon to be analyzed in the prediction period based on the occurrence probabilities of all intermediate events and the top event and the decision guidance comprehensive performance values of each level in the fault tree.

[0023] Preferably, the fault probability evaluation submodule comprises:

[0024] The first fault probability evaluation unit is configured to perform piecewise modeling based on the wear amount threshold of each automobile component and the wear trend data of the automobile component to obtain a piecewise failure rate function of each automobile component, and determine first fault probability data of each automobile component in the prediction period based on the piecewise failure rate function of each automobile component.

[0025] The second fault probability evaluation unit is configured to determine second fault probability data of each automobile component in the prediction period based on a future wear feature of each automobile component and an association model between the failure rate and the wear feature.

[0026] a fault probability synthesizing unit, configured to time-align and weight-sum the first fault probability data and the second fault probability data of each automobile component in the prediction period, to obtain the final fault probability data of each automobile component in the prediction period;

[0027] an event probability determining unit, configured to determine the occurrence probability time series data of all basic events in the prediction period based on the final fault probability data of each automobile component in the prediction period.

[0028] Preferably, the fault synchronism value evaluation submodule comprises:

[0029] a first synchronism value setting unit, configured to set the fault synchronism value of the corresponding minimal cut set as 1 when the single minimal cut set of the top event only contains one basic event;

[0030] a same-order matrix building unit, configured to, when the single minimal cut set of the top event contains more than one basic event, synchronize and equally divide the occurrence probability time series data of all basic events in the corresponding minimal cut set, to obtain all occurrence probability sets of each basic event, and build the occurrence probability same-order matrix of each basic event in the corresponding minimal cut set based on all occurrence probability sets of all basic events in the corresponding minimal cut set;

[0031] a feature vector generating unit, configured to perform eigenvalue decomposition on each occurrence probability same-order matrix, to obtain all eigenvalues of each occurrence probability same-order matrix, and to sort all eigenvalues of each occurrence probability same-order matrix from large to small, to obtain the eigenvalue vector of each occurrence probability same-order matrix;

[0032] a second synchronism value setting unit, configured to correspondingly calculate the similarity of all same-row difference value vectors, all same-column difference value vectors and the eigenvalue vector of the occurrence probability same-order matrix of all basic events in the single minimal cut set, to obtain multiple similarities of all basic events in the corresponding minimal cut set, and to determine the fault synchronism value of the corresponding minimal cut set based on all similarities of all basic events in the single minimal cut set.

[0033] Preferably, the guide performance evaluation submodule comprises:

[0034] an iterative synchronism value determining unit, configured to determine all common upper nodes of each minimal cut set in the fault tree, to determine the iterative synchronism value of each minimal cut set at the corresponding each common upper node based on the belonging level of all basic events contained in each minimal cut set in the fault tree, the level difference between the corresponding each common upper node and the fault synchronism value of the corresponding minimal cut set.

[0035] A relative layer height ratio determining unit is configured to determine a relative layer height ratio of each common upper layer node of each minimum essential cut set in all common upper layer nodes of the corresponding minimum essential cut set;

[0036] A decision guidance comprehensive performance determining unit is configured to determine a decision guidance comprehensive performance value of each level in the fault tree based on the iteration synchronization value of all common upper layer nodes contained in each level in the fault tree and the relative layer height ratio of all common upper layer nodes of the corresponding minimum essential cut set.

[0037] Preferably, the hidden danger prediction evaluation sub-module comprises:

[0038] A guidance occurrence probability determining unit is configured to determine a guidance occurrence probability of all intermediate events and top events based on the occurrence probability of all intermediate events and top events and the decision guidance comprehensive performance value of each level in the fault tree.

[0039] A hidden danger prediction evaluation unit is configured to take all intermediate events and / or top events whose guidance occurrence probability exceeds a threshold value in the fault tree of the fault phenomenon to be analyzed as a relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in a prediction period within a range of levels involved in the corresponding fault tree.

[0040] Preferably, the maintenance strategy generation module comprises:

[0041] A maintenance item list retrieval sub-module is configured to determine a highest level and a lowest level in the relative hidden danger level and a total number of covered levels, retrieve a maintenance item detail list corresponding to the fault phenomenon to be analyzed based on the highest level and the lowest level in the relative hidden danger level and the total number of covered levels, and determine all maintenance items and corresponding maintenance plan details.

[0042] A maintenance strategy generation sub-module is configured to generate a maintenance strategy based on all maintenance items and corresponding maintenance plan details.

[0043] The present application provides an automobile fault diagnosis and predictive maintenance method based on a fault tree analysis method, which comprises:

[0044] S1: constructing a fault tree of a fault phenomenon to be analyzed based on a top event and all intermediate events and all basic events of the fault phenomenon to be analyzed;

[0045] S2: determining all minimum essential cut sets of the top event based on the fault tree of the fault phenomenon to be analyzed;

[0046] S3: calculating occurrence probability time sequence data of all intermediate events and top events based on the loss trend data of automobile components, and analyzing a relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in a prediction period in combination with all minimum essential cut sets of the top event;

[0047] S4: generating a maintenance strategy based on a fault tree of the fault phenomenon to be analyzed at a relative hidden danger level of the prediction period.

[0048] The beneficial effects generated by the present application relative to the prior art are: the fault tree construction module constructs a fault tree based on top events, intermediate events and basic events, providing a clear logical framework for the entire fault analysis, intuitively presenting the hierarchical relationship between fault causes, comprehensively analyzing the logical relationship between fault phenomena and events at each level, and facilitating a comprehensive and systematic understanding of the fault structure. The minimum cut set determination module accurately locates the key factor combination causing the fault by determining the minimum necessary cut set of the top event, focusing on the core points for subsequent analysis. Furthermore, the hidden danger prediction and evaluation module calculates event occurrence probability time series data in combination with automobile component wear trend data, and analyzes the relative hidden danger level in combination with the minimum necessary cut set, realizes dynamic evaluation and prediction of fault hidden dangers, perceives potential risks in advance, and greatly improves the accuracy and reliability of the prediction results. The maintenance strategy generation module generates a scientific and reasonable maintenance strategy according to the relative hidden danger level, making the maintenance work more targeted and forward-looking, reasonably arranging maintenance resources and timing, effectively preventing faults, ensuring the safety of automobile operation, reducing maintenance costs, and improving the overall reliability and service life of the automobile.

[0049] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application can be realized and attained by the structure particularly pointed out in the written description and claims.

[0050] The technical solutions of the present application will be further described in detail below with the help of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0051] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation on the present application. In the drawings:

[0052] Figure 1 The schematic diagram of the automobile fault diagnosis and prediction maintenance system based on the fault tree analysis method in the embodiment of the present application;

[0053] Figure 2 The schematic diagram of the fault tree construction module in the embodiment of the present application;

[0054] Figure 3 The schematic diagram of the minimum cut set determination module in the embodiment of the present application;

[0055] Figure 4 The schematic diagram of the hidden danger prediction and evaluation module in the embodiment of the present application;

[0056] Figure 5A schematic diagram of a maintenance strategy generation module in an embodiment of the present application. DETAILED DESCRIPTION

[0057] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings, in which it is understood that the preferred embodiments described below are merely meant to illustrate and explain the present application, and are not meant to limit the present application.

[0058] Embodiment 1

[0059] The present application provides an automobile fault diagnosis and predictive maintenance system based on a fault tree analysis method, referring to Figure 1 , comprising:

[0060] a fault tree construction module, configured to construct a fault tree of a fault phenomenon to be analyzed based on a top event of the fault phenomenon to be analyzed and all intermediate events and all basic events;

[0061] a minimum cut set determination module, configured to determine all minimum necessary cut sets of the top event based on the fault tree of the fault phenomenon to be analyzed;

[0062] a hidden danger prediction and evaluation module, configured to calculate occurrence probability time sequence data of all intermediate events and the top event based on loss trend data of automobile components, and analyze a relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in a prediction period in combination with all minimum necessary cut sets of the top event;

[0063] a maintenance strategy generation module, configured to generate a maintenance strategy based on the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period.

[0064] In this embodiment, the fault phenomenon to be analyzed refers to a fault situation that needs to be diagnosed and predictively maintained by the system in automobile operation. For example, difficult starting of an automobile engine, body shaking of a vehicle during driving, etc. are starting points of subsequent fault analysis work.

[0065] In this embodiment, the top event of the fault phenomenon to be analyzed is an event representing a final presented state of a fault, which is at the top end of the fault tree; the intermediate event is between the top event and the basic event, which is caused by the basic event and has an impact on the top event; and the basic event is the bottom layer of the fault tree, which cannot be further subdivided and is the cause of the fault. For example, for the automobile "abnormal light" fault, "abnormal light" is the top event, "light bulb fault", "line short circuit", etc. can be used as the intermediate event, and "light bulb filament fuse", "line damage", etc. belong to the basic event.

[0066] In this embodiment, the fault tree is a tree structure built based on the top event, intermediate event and basic event of the fault phenomenon to be analyzed, according to the logical relationship. Through the logical gate to connect each event, the fault formation logical path from the basic event to the intermediate event and then to the top event is displayed, for example, the logical relationship such as "and gate" and "or gate" reflects the effect of event combination on fault occurrence.

[0067] In this embodiment, the wear trend data of the automobile parts is the data about the change of the wear of each part of the automobile with time or the number of uses and other factors during use. For example, the data record of the thickness of the brake pad decreasing with the number of braking, or the data of the performance of the engine oil decreasing with the mileage.

[0068] In this embodiment, the occurrence probability time series data of the intermediate event and the top event is a data sequence composed of the probabilities of the occurrence of the intermediate event and the top event at different time points in the prediction period, which is calculated based on the wear trend data of the automobile parts. For example, the sequence of the occurrence probability of the intermediate event "brake failure" every day in the next month.

[0069] In this embodiment, the prediction period is a time period set in advance for the evaluation of the automobile fault related events. In this time period, the analysis of the occurrence probability of each event of the fault tree based on the wear trend data of the automobile parts is carried out. For example, it is set to be one week in the future.

[0070] Embodiment 2:

[0071] On the basis of embodiment 1, the fault tree construction module refers to Figure 2 , including:

[0072] The event determination sub-module is used to determine the top event and all intermediate events and all basic events based on the automobile operation related data of the fault phenomenon to be analyzed;

[0073] The fault tree building sub-module is used to build a tree structure for the top event and all intermediate events and all basic events of the fault tree of the fault phenomenon to be analyzed based on the logical relationship between all events, to obtain the fault tree of the fault phenomenon to be analyzed.

[0074] In this embodiment, the automobile operation related data of the fault phenomenon to be analyzed refers to various data information generated or collected during the operation of the automobile. These data include but are not limited to the working parameters of each part of the automobile, such as the engine speed, temperature, oil pressure, and the vehicle speed, mileage, etc.; the data feedback by various sensors, such as the data of oxygen sensor and wheel speed sensor; in addition, the fault codes recorded by the automobile electronic control system, and the environmental data of the vehicle, such as temperature, humidity, altitude, etc., all belong to this category of data.

[0075] In this embodiment, the top event and all intermediate events and all basic events are determined based on the vehicle operation related data of the fault phenomenon to be analyzed, that is, by means of the above vehicle operation related data, the different level events causing the fault phenomenon to be analyzed are found out through analysis and reasoning. The top event is determined from the fault phenomenon, for example, the fault of difficult starting of the vehicle, and the top event is "difficult starting of the vehicle". Then, it is judged from the data which intermediate link faults can cause the top event as the intermediate event, for example, "insufficient battery power", "ignition system fault" and the like. At the same time, the root causes causing these intermediate events are mined from the data as the basic events, such as "battery aging", "spark plug damage" and the like.

[0076] In this embodiment, the fault tree top event and all intermediate events and all basic events of the fault phenomenon to be analyzed are built in tree structure based on the logical relationship between all events, and the fault tree of the fault phenomenon to be analyzed is obtained. The top event is placed at the top end of the fault tree according to the cause-effect association between events. Then, the intermediate events and the basic events are arranged in turn below the top event according to the logical order of causing the fault. The events are connected by "and gate", "or gate" and the like to build a tree structure. For example, if "insufficient battery power" and "ignition system fault" can only cause "difficult starting of the vehicle", then the two intermediate events and the top event are connected by "and gate"; if either "spark plug damage" or "ignition coil fault" can cause "ignition system fault", then they are connected with "or gate" with "ignition system fault", and finally the fault tree is formed.

[0077] The beneficial effects of the above technology are that the event determination sub-module determines the top event, the intermediate event and the basic event based on the vehicle operation related data, uses the actual operation data to make the determined events more consistent with the real fault situation of the vehicle, provides an accurate basis for subsequent fault analysis, and improves the pertinence and reliability of fault diagnosis. The fault tree building sub-module builds a tree structure according to the logical relationship between events, can clearly and intuitively present the cause-effect association between events, displays the generation path of the fault in a graphical way, facilitates technicians to understand and analyze complex faults, provides a good framework support for subsequent minimum necessary cut set determination, hidden danger prediction and evaluation, and makes the whole fault diagnosis and prediction maintenance system run more scientifically and efficiently, which helps to more accurately locate the fault cause, develop maintenance strategies in advance, ensure stable operation of the vehicle, and reduce the probability of fault occurrence and maintenance cost.

[0078] Embodiment 3:

[0079] On the basis of embodiment 1, the minimum cut set determination module refers to Figure 3 , and includes:

[0080] The top event expression construction submodule is configured to represent all basic events in the fault tree by Boolean variables, and to perform logical deduction step by step upwards in the fault tree to obtain a Boolean expression of the top event.

[0081] The expression simplification and expansion submodule is configured to expand the Boolean expression of the top event based on preset simplification rules to obtain a plurality of basic event product terms, and to regard each basic event product term as a single minimal necessary cut set of the top event.

[0082] In this embodiment, the Boolean expression of the top event is obtained by representing all basic events in the fault tree by Boolean variables and performing logical deduction step by step upwards in the fault tree. That is, a Boolean variable (usually represented by a letter and taking a value of 0 or 1, where 0 represents that an event does not occur and 1 represents that an event occurs) is set for each basic event in the fault tree. Then, according to the logical relationship (such as the relationship represented by logical gates such as AND and OR) between events in the fault tree, the Boolean expression of the top event is obtained by performing logical deduction step by step upwards in the order of levels, starting from the basic events. For example, if an intermediate event is caused by two basic events through an AND gate, the Boolean expression of the intermediate event is the product of the Boolean variables of the two basic events. If the intermediate event is caused by two basic events through an OR gate, the Boolean expression of the intermediate event is the sum (logical sum, that is, 1 as long as one is 1) of the Boolean variables of the two basic events. The Boolean expression of the top event is obtained by performing such layer-by-layer deduction, and the expression reflects the logical relationship between the top event and the basic events.

[0083] In this embodiment, the preset simplification rules are a series of criteria for simplifying the Boolean expression, which are set in advance. These rules are usually based on the basic operation laws of Boolean algebra, such as commutative law (A+B=B+A, AB=BA), associative law ((A+B)+C=A+(B+C), (AB)C=A(BC)), distributive law (A(B+C)=AB+AC), absorption law (A+AB=A, A(A+B)=A), etc. Some rules may also be developed for specific fault tree analysis scenarios, with the purpose of simplifying complex Boolean expressions to facilitate clearer analysis and understanding of the logical relationship represented by the expression.

[0084] In this embodiment, the Boolean expression of the top event is expanded based on the preset simplification rules to obtain a plurality of basic event product terms. That is, the obtained Boolean expression of the top event is processed by using the above-mentioned preset simplification rules. The expression is expanded step by step by using the commutative law, the distributive law, etc. For example, for the Boolean expression A(B+C), AB+AC is obtained according to the distributive law, so that the originally complex expression is expanded into the form of a plurality of basic event product terms.

[0085] The beneficial effects of the above technology are that the top event expression construction submodule represents the basic events in the fault tree with Boolean variables and derives the Boolean expression of the top event through step-by-step upward logical deduction. This way converts the complex logical relationship of the fault tree into a simple mathematical expression, making the fault analysis process more logical and systematic, facilitating understanding and operation, and laying a foundation for subsequent accurate determination of the minimum necessary cut sets. The expression simplification expansion submodule expands the top event Boolean expression according to the preset simplification rules, obtains multiple basic event product terms and regards them as a single minimum necessary cut set of the top event, and through this way can accurately locate the minimum basic event combination causing the top event to occur, helping maintenance personnel to accurately judge the key factors causing the fault, avoiding blind search in fault troubleshooting, improving the fault diagnosis efficiency, making the maintenance resources more reasonably allocated, and thus effectively improving the overall performance of the automobile fault diagnosis and predictive maintenance system based on the fault tree analysis method.

[0086] Embodiment 4:

[0087] On the basis of embodiment 1, the hidden danger prediction and evaluation module refers to Figure 4 , and comprises:

[0088] The fault probability evaluation submodule is configured to determine occurrence probability time sequence data of all basic events in a prediction period based on the wear trend data of the automobile components;

[0089] The event probability derivation submodule is configured to derive occurrence probability time sequence data of all intermediate events and the top event in the prediction period based on the fault tree;

[0090] The fault synchronism value evaluation submodule is configured to analyze fault synchronism values of all minimum necessary cut sets of the top event based on the occurrence probability time sequence data of all basic events;

[0091] The guidance performance evaluation submodule is configured to assign values to each level in the fault tree based on the fault synchronism values of all minimum necessary cut sets of the top event, and obtain decision guidance comprehensive performance values of each level in the fault tree;

[0092] The hidden danger prediction and evaluation submodule is configured to determine the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period based on the occurrence probability of all intermediate events and the top event and the decision guidance comprehensive performance values of each level in the fault tree.

[0093] In this embodiment, the occurrence probability time series data of all basic events in the prediction period is determined based on the wear trend data of the automobile parts, i.e., by analyzing the data of the wear change of each part of the automobile over time or the number of times of use, etc., to calculate the probability of failure of each basic event at different time points in a pre-set prediction period, and form a corresponding data sequence. For example, according to the wear trend data of the wear amount of the engine piston ring with the increase of the driving mileage, combined with the related failure probability model, the probability of failure of the related basic event (such as piston ring poor sealing) caused by the wear of the piston ring every day in the future one month (prediction period) is determined, so as to obtain the occurrence probability time series data of the basic event in the prediction period.

[0094] In this embodiment, the occurrence probability time series data of all intermediate events and top events in the prediction period is obtained by upward derivation of the occurrence probability time series data of all basic events in the prediction period based on the fault tree, which is derived from the occurrence probability time series data of the basic events in the prediction period according to the logical relationship between the events in the fault tree, and gradually calculates the occurrence probability of the intermediate events and the top events at different time points in the same prediction period, and then obtains the corresponding data sequence. For example, if an intermediate event is caused by two basic events through an "and gate" logical relationship, then the occurrence probability of the intermediate event at a time is the product of the occurrence probabilities of the two basic events at the same time; if through an "or gate" logical relationship, the occurrence probability of the intermediate event at a time is the sum of the occurrence probabilities of the two basic events at the same time minus the probability of their simultaneous occurrence (to avoid repeated calculation). According to such logical relationship, the occurrence probability time series data of all intermediate events and top events in the prediction period can be obtained by upward derivation along the fault tree from bottom to top.

[0095] In this embodiment, the decision guidance comprehensive performance value of each level in the fault tree is a value given to each level in the fault tree based on the fault synchronism of all minimal cut sets of the top event. It comprehensively considers the related information of the minimal cut sets in each level, such as the belonging level of the basic events contained in the minimal cut set in the fault tree, the level difference with the common upper node, and the fault synchronism value of the minimal cut set, etc. This value is used to measure the importance of each level in the fault tree in the process of fault diagnosis and prediction and the guiding role of decision making. The higher the value, the more critical the level may be in the fault analysis and decision making. Maintenance personnel can pay more attention to and investigate the events in different levels according to these values, so as to improve the efficiency and accuracy of fault diagnosis and prediction.

[0096] The above technology has the beneficial effects that: the fault probability evaluation submodule determines the occurrence probability time series data of the basic event in the prediction period based on the automobile component wear trend data, uses the actual wear condition of the component to make probability prediction, so that the result is closer to the actual situation, and accurate data basis is provided for subsequent analysis. The event probability derivation submodule derives the occurrence probability time series data of the intermediate event and the top event according to the fault tree, and builds a complete probability chain from the basic event to the high-level event, and comprehensively presents the transmission process of the fault occurrence possibility. The fault synchronicity value evaluation submodule analyzes the fault synchronicity value of the minimum necessary cut set, can understand the possibility of multiple basic events cooperating to cause a fault, and is helpful for discovering potential fault modes. The guidance performance evaluation submodule values each level of the fault tree to obtain a decision guidance comprehensive performance value, and quantifies the importance of different levels in fault diagnosis and prediction. Finally, the hidden danger prediction evaluation submodule determines the relative hidden danger level based on the above information, so that the maintenance personnel can clearly understand the fault hidden danger degree, plan the maintenance strategy in advance, accurately allocate maintenance resources, improve the scientificity and effectiveness of automobile fault diagnosis and predictive maintenance, and ensure the safe and reliable operation of the automobile.

[0097] Embodiment 5:

[0098] On the basis of embodiment 4, the fault probability evaluation submodule refers to Figure 4 , and includes:

[0099] The first fault probability evaluation unit is configured to perform segmented modeling based on the wear amount threshold of each automobile component and the wear trend data of the automobile component, obtain a segmented failure rate function of each automobile component, and determine first fault probability data of each automobile component in the prediction period based on the segmented failure rate function of each automobile component.

[0100] The second fault probability evaluation unit is configured to determine second fault probability data of each automobile component in the prediction period based on the future wear characteristics of each automobile component and a correlation model between the failure rate and the wear characteristics.

[0101] The fault probability comprehensive unit is configured to time-align and weight-add the first fault probability data and the second fault probability data of each automobile component in the prediction period to obtain final fault probability data of each automobile component in the prediction period.

[0102] The event probability determination unit is configured to determine occurrence probability time series data of all basic events in the prediction period based on the final fault probability data of each automobile component in the prediction period.

[0103] In this embodiment, the wear amount division threshold of each automobile component is a pre-set wear amount limit value for each component of the automobile. Through these limit values, the wear condition of the component can be divided into different stages. For example, for the automobile brake pad, several threshold values can be set according to its thickness, such as dividing the brake pad thickness from the initial value to the wear to the replacement standard value into stages of light wear, moderate wear, severe wear, etc., and the limit values of each stage are the wear amount division threshold values, which are used for subsequent definition of the wear stage of the component.

[0104] In this embodiment, the segmented modeling is performed based on the wear amount division threshold of each automobile component and the wear trend data of the automobile component, the segmented failure rate function of each automobile component is obtained, and the first failure probability data of each automobile component in the prediction period is determined based on the segmented failure rate function of each automobile component. The specific process is as follows: according to the wear amount division threshold, the component wear trend data is divided according to different wear stages. The data of each stage is analyzed, and a suitable mathematical model (such as regression analysis method) is used to construct the failure rate function corresponding to each stage, i.e. the segmented failure rate function. These functions describe the relationship between the probability of component failure and time or other related variables in different wear stages. Then, according to the wear stage of the component in the prediction period, the corresponding segmented failure rate function is substituted to calculate the failure probability of the component at different time points in the prediction period, so as to determine the first failure probability data of each automobile component in the prediction period.

[0105] In this embodiment, the future wear characteristics of each automobile component refer to the performance values of the related quantities (such as temperature, load, oil contamination degree or piston ring wear amount, etc.) that will cause the wear of the automobile component in the future prediction period.

[0106] In this embodiment, the correlation model of failure rate and wear characteristics is a mathematical relationship model established by research and analysis, which is used to describe the relationship between the failure rate of the automobile component and its various wear characteristics. This model can be obtained based on a large amount of experimental data, actual use records and theoretical analysis. For example, through the tracking research of a large number of same type automobile components, it is found that the failure rate of the component and the wear characteristics such as temperature, load, oil contamination degree or piston ring wear amount have a certain functional relationship, and this relationship is expressed by a mathematical expression, which forms the correlation model of failure rate and wear characteristics, which is used for subsequent prediction of the failure rate of the component according to its wear characteristics.

[0107] In this embodiment, the second failure probability data of each automobile component in the prediction period is determined based on the future wear characteristics of each automobile component and the correlation model between failure rate and wear characteristics, that is, the future wear characteristic information of each automobile component in the prediction period is substituted into the established correlation model between failure rate and wear characteristics. The model calculates the failure probability of the component at different time points in the prediction period according to the input wear characteristic data, and these probability data constitute the second failure probability data of each automobile component in the prediction period.

[0108] In this embodiment, the first failure probability data and the second failure probability data of each automobile component in the prediction period are time-aligned and weightedly added to obtain the final failure probability data of each automobile component in the prediction period, that is, the first failure probability data based on the piecewise modeling and the second failure probability data based on the correlation model are matched according to time sequence (time alignment) so that they correspond to each other in the time dimension. Then, according to the reliability, importance and other factors of the two data sources, certain weights are respectively assigned to them (for example, the weights of the first failure probability data and the second failure probability data are both 0.5). The two failure probability data at the corresponding time points are respectively multiplied by the respective weights and then added to obtain the final failure probability at each time point, and all the final failure probabilities at the time points constitute the final failure probability data of each automobile component in the prediction period.

[0109] In this embodiment, the occurrence probability time series data of all basic events in the prediction period is determined based on the final failure probability data of each automobile component in the prediction period. Since the basic events of the automobile are often associated with specific automobile components, the final failure probability data of each automobile component reflects the possibility of failure of the component at different time points in the prediction period. For a basic event triggered by failure of a specific automobile component, the occurrence probability time series data of the basic event in the prediction period can directly refer to the final failure probability data of the component. For example, if a basic event is caused by failure of a component of a specific engine, the occurrence probability time series data of the basic event in the prediction period corresponds to the final failure probability data of the engine component in the prediction period, thereby determining the occurrence probability time series data of all basic events in the prediction period.

[0110] The beneficial effects of the above technology are: the first failure probability evaluation unit divides the threshold based on the component wear amount and the wear trend data segment modeling, obtains the segmented failure rate function and determines the first failure probability data. This segmented modeling based on actual wear can more accurately reflect the failure probability of components at different wear stages and fit the real aging process of components. The second failure probability evaluation unit determines the second failure probability data according to the future wear characteristics of the component and the failure rate and wear characteristics correlation model, considers the potential failure risk of the component from another dimension, and provides a supplementary perspective for failure probability evaluation. The failure probability synthesis unit time-aligns the two types of failure probability data and obtains the final failure probability data by weighted addition, which combines the advantages of the two evaluation methods, making the result more comprehensive and accurate. The event probability determination unit determines the basic event occurrence probability time series data based on the final failure probability data, provides a reliable data foundation for subsequent upward derivation based on the fault tree, enables the hidden danger prediction and evaluation module to more accurately analyze the fault hidden danger, helps maintenance personnel to more scientifically judge the fault possibility, and plans maintenance measures in advance, improves the accuracy and practicality of the automobile fault diagnosis and predictive maintenance system, and ensures the stable operation of the automobile.

[0111] Embodiment 6:

[0112] On the basis of embodiment 4, the fault synchronicity value evaluation sub-module, referring to Figure 4 , includes:

[0113] The first synchronicity value setting unit is configured to, when the single minimum essential cut set of the top event contains only one basic event, set the fault synchronicity value of the corresponding minimum essential cut set to 1.

[0114] The same sequence matrix building unit is configured to, when the single minimum essential cut set of the top event contains more than one basic event, synchronize and evenly divide the occurrence probability time series data of all basic events in the corresponding minimum essential cut set, and synchronize and disorder the order of the occurrence probability time series data of all basic events in the corresponding minimum essential cut set, to obtain all occurrence probability sets of each basic event in the corresponding minimum essential cut set, and build an occurrence probability same sequence matrix of each basic event in the corresponding minimum essential cut set based on all occurrence probability sets of all basic events in the corresponding minimum essential cut set.

[0115] The eigenvector generation unit is configured to perform eigenvalue decomposition on each occurrence probability same sequence matrix to obtain all eigenvalues of each occurrence probability same sequence matrix, and sort all eigenvalues of each occurrence probability same sequence matrix from large to small to obtain an eigenvector of each occurrence probability same sequence matrix.

[0116] The second synchronism value setting unit is configured to calculate the similarity of all the same-row difference value vectors, all the same-column difference value vectors and the eigenvalue vector of the same-sequence matrix of the occurrence probability of all the basic events in the single minimum necessary cut set, to obtain a plurality of similarities of all the basic events in the corresponding minimum necessary cut set, and to determine the fault synchronism value of the corresponding minimum necessary cut set based on all the similarities of all the basic events in the single minimum necessary cut set.

[0117] In this embodiment, the occurrence probability time series data of all the basic events in the corresponding minimum necessary cut set are synchronously divided and synchronously shuffled to obtain all the occurrence probability sets of each basic event, and the same-sequence matrix of the occurrence probability of each basic event in the corresponding minimum necessary cut set is built based on all the occurrence probability sets of all the basic events in the corresponding minimum necessary cut set. Specifically, the synchronous division is to divide the occurrence probability time series data of each basic event according to the same time interval or other unified standard, so that the data of different basic events are consistent in the time scale. The synchronous shuffling is to synchronously and randomly shuffle the order of each segment of data of different basic events after division, for example, to exchange the value originally in the first position with the value originally in the third position. This is to improve the generalization ability of the same-sequence matrix of the occurrence probability, so as to more comprehensively analyze the relationship between events. Through these operations, a series of occurrence probability values of each basic event, i.e. all the occurrence probability sets, are obtained. Then, based on these sets, the probability sets of each basic event in the corresponding minimum necessary cut set are arranged into a matrix form according to certain rules, and the same-sequence matrix of the occurrence probability is built.

[0118] In this embodiment, the same-row difference value vector refers to the vector formed by calculating the difference between adjacent elements in each row of the same-sequence matrix of the occurrence probability. For example, for a row element [a1, a2, a3], the same-row difference value vector is [a2-a1, a3-a2].

[0119] In this embodiment, the same-column difference value vector refers to the vector obtained by calculating the difference between adjacent elements in each column of the same-sequence matrix of the occurrence probability. For example, for a column element [b1, b2, b3], the same-column difference value vector is [b2-b1, b3-b2].

[0120] In this embodiment, the similarity of all the same row difference vectors, all the same column difference vectors, and the eigenvalue vectors of the same order matrix of the occurrence probability of all basic events in a single minimum cut set is calculated, so as to measure the similarity between the vectors from multiple dimensions, and then evaluate the fault synchronism between the basic events. The similarity calculation here can use some common similarity measurement methods, such as cosine similarity, Euclidean distance similarity, etc. For the same row difference vectors, the similarity between the difference vectors of the corresponding rows in the same order matrix of different basic events is calculated, so as to understand the similarity between different basic events under the same probability change trend of the same basic event; the same operation is performed on the same column difference vectors, so as to analyze the synchronization of the probability change of different basic events from the column angle; the eigenvalue vector contains important feature information of the matrix, and by calculating the similarity of the eigenvalue vector, the similarity between the basic events can be evaluated from the overall structure level. The similarity of these different vectors is comprehensively considered, so that the relationship between the basic events can be more comprehensively reflected.

[0121] In this embodiment, the fault synchronism value of the corresponding minimum cut set is determined based on all the similarities of all the basic events in a single minimum cut set, that is, all the same row difference vector similarities, the same column difference vector similarities, and the eigenvalue vector similarities are comprehensively considered, and a predetermined method (such as weighted average) is used for summary calculation, and finally a numerical value is obtained, which is the fault synchronism value of the corresponding minimum cut set. The value is used to represent the synchronization degree of all the basic events in the minimum cut set in the probability change of fault occurrence, and the higher the value, the more synchronized the probability change of fault occurrence of these basic events, that is, they are more likely to cause faults at the same time, which provides an important reference for fault hidden danger evaluation.

[0122] The beneficial effects of the above technology are: when the minimum cut set contains multiple basic events, the same order matrix building unit synchronously divides and randomly arranges the time sequence data of the occurrence probability of the basic events, builds the occurrence probability same order matrix, and comprehensively captures the complex probability correlation between events. The feature vector generation unit performs eigenvalue decomposition on the same order matrix and sorts the eigenvalue vector to effectively extract key information. The second synchronism value setting unit determines the fault synchronism value by calculating the similarity of the same row difference vectors, the same column difference vectors, and the eigenvalue vectors, comprehensively considers from multiple dimensions, and ensures that the evaluation result is comprehensive and accurate. These steps work together to help maintenance personnel deeply understand the fault occurrence mode, plan maintenance strategies in advance, and optimize the automobile fault diagnosis and predictive maintenance system based on the fault tree analysis method.

[0123] Embodiment 7:

[0124] On the basis of embodiment 4, a performance evaluation sub-module is guided, and reference is made to Figure 4 , which includes:

[0125] an iterative synchronicity value determination unit configured to determine an iterative synchronicity value of each minimal cut set at each common upper node of the minimal cut set based on a level difference between a level of each basic event included in the minimal cut set and each common upper node and a fault synchronicity value of the minimal cut set;

[0126] a relative level ratio determination unit configured to determine a relative level ratio of each common upper node of each minimal cut set among all common upper nodes of the minimal cut set;

[0127] a decision guidance comprehensive performance determination unit configured to determine a decision guidance comprehensive performance value of each level in the fault tree based on the iterative synchronicity value of each common upper node included in each level and the relative level ratio among all common upper nodes of the minimal cut set corresponding to each level.

[0128] In this embodiment, determining all common upper nodes of each minimal cut set in the fault tree means finding, for each minimal cut set, those nodes in the tree structure above it and connected to all basic events in the minimal cut set. These nodes represent higher-level fault events or intermediate fault states caused by the minimal cut set. For example, in a fault tree describing the faults of a car engine, a minimal cut set includes "spark plug failure" and "ignition coil failure", and "ignition system failure" can be a common upper node of the minimal cut set because both basic events are related to "ignition system failure" and below it in the level.

[0129] In this embodiment, the iterative synchronicity value of each minimal cut set at each common upper node is determined based on a level difference between a level of each basic event included in the minimal cut set and each common upper node and a fault synchronicity value of the minimal cut set. Specifically, the level of the basic event in the minimal cut set and the level of each common upper node are first determined to obtain the level difference, which reflects the logical distance from the basic event to the common upper node. Then, the iterative synchronicity value of each minimal cut set at each common upper node is calculated by multiplying the ratio of the level difference to the total number of levels in the fault tree and the fault synchronicity value of the minimal cut set. This value takes into account the level relationship between the basic event and the upper node and the fault synchronicity between the basic events, and is used to measure the influence degree and the synchronous correlation degree of the minimal cut set on the occurrence of a specific common upper node fault.

[0130] In this embodiment, the relative layer height ratio of each common upper node of each minimum necessary cut set in all common upper nodes of the corresponding minimum necessary cut set is determined, that is, for all common upper nodes of a certain minimum necessary cut set, the height proportion of each common upper node relative to other common upper nodes is calculated respectively. For example, if a minimum necessary cut set has three common upper nodes A, B and C, node A is 3 layers away from the layer level of the basic event, node B is 5 layers away, and node C is 4 layers away. Taking the layer level number 5 of the node B farthest from the basic event layer level as the denominator, the relative layer height ratio of node A is calculated as 3 ÷ 5 = 0.6, and the relative layer height ratio of node C is calculated as 4 ÷ 5 = 0.8. The relative layer height ratio reflects the relative importance or influence weight of each common upper node in all common upper nodes. The higher the layer level, the greater the relative layer height ratio, and the greater the influence on fault propagation and overall fault condition.

[0131] In this embodiment, based on the iteration synchronization value of all common upper nodes contained in each level of the fault tree and the relative layer height ratio in all common upper nodes of the corresponding minimum necessary cut set, the decision guidance comprehensive performance value of each level in the fault tree is determined. The specific method is to take the ratio of the relative layer height ratio of each common upper node contained in each level in all common upper nodes of the corresponding minimum necessary cut set to the relative layer height ratio of all common upper nodes contained in the corresponding level in all common upper nodes of the corresponding minimum necessary cut set as the weight of the corresponding common upper node contained in the corresponding level, and to add the iteration synchronization values of all common upper nodes contained in the corresponding level based on the weights of all common upper nodes contained in each level to obtain the decision guidance comprehensive performance value of each level. This value represents the comprehensive importance and guidance value of the level in the entire fault tree fault diagnosis and prediction process. The higher the value, the more critical the level in fault analysis and decision making, which can help maintenance personnel pay more attention to specific levels of events when troubleshooting and predicting maintenance.

[0132] The above-mentioned technical beneficial effects are: the iteration synchronization value determination unit accurately evaluates the relationship with the upper node by comprehensively considering the minimum necessary cut set basic event belonging level, the layer level difference with the common upper node, and the fault synchronization value, and deeply reflects the fault transmission and cooperation situation; the relative layer height ratio determination unit quantifies the layer level weight by clearly determining the relative layer height ratio of each common upper node, and reflects the importance difference of different levels; the decision guidance comprehensive performance determination unit scientifically determines the decision guidance comprehensive performance value of each level by combining the above two, provides clear guidance for fault diagnosis, so that maintenance personnel can target troubleshooting and pay more attention to key levels, improve fault diagnosis efficiency and accuracy, and further improve the system function based on the fault tree analysis method.

[0133] Embodiment 8:

[0134] On the basis of Embodiment 4, the hidden danger prediction and evaluation sub-module, referring to Figure 4 , comprises:

[0135] a guidance occurrence probability determination unit, configured to determine the guidance occurrence probability of all intermediate events and top events based on the occurrence probability of all intermediate events and top events and the decision guidance comprehensive performance value of each level in the fault tree;

[0136] a hidden danger prediction and evaluation unit, configured to regard all intermediate events and / or top events whose guidance occurrence probability exceeds the threshold value in the fault tree of the fault phenomenon to be analyzed as the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period within the range of the level involved in the corresponding fault tree.

[0137] In this embodiment, the guidance occurrence probability of all intermediate events and top events is determined based on the occurrence probability of all intermediate events and top events and the decision guidance comprehensive performance value of each level in the fault tree. This means that the calculation method of multiplying the probability of failure of each intermediate event and top event in the fault tree and the decision guidance comprehensive performance value of the level they belong to (the occurrence probability and the decision guidance comprehensive performance value are multiplied) is considered comprehensively to obtain a new probability value, i.e. the guidance occurrence probability. This probability value not only reflects the possibility of the occurrence of the event itself, but also incorporates the influence of the level on fault analysis and decision-making, and more comprehensively reflects the actual importance of the event in the entire fault system and the potential guiding role of the event in the occurrence of the fault.

[0138] In this embodiment, the threshold value is a pre-set standard value. When evaluating the hidden danger of the fault, it serves as a reference limit for judging whether the guidance occurrence probability of the intermediate event and the top event is at a level that needs to be concerned. The setting of this threshold value is usually based on experience, historical data or the acceptable degree of automobile fault risk and other factors. For example, if the threshold value is set to 0.6, then the intermediate event or top event with a guidance occurrence probability greater than 0.6 can be regarded as an object that needs to be focused on, because its occurrence has a greater impact on the overall fault and may imply a higher hidden danger of the fault.

[0139] In this embodiment, the level range involved in the corresponding fault tree of all intermediate events and / or top events whose leading occurrence probability exceeds the threshold refers to, in the fault tree structure, when the leading occurrence probability of certain intermediate events and top events is determined to exceed the pre-set threshold, finding the level of these events in the fault tree and explicitly specifying the range covered by these levels. For example, if the leading occurrence probability of "engine overheating" (top event) and "coolant leakage" (intermediate event) exceeds the threshold, "engine overheating" is at the 3rd level of the fault tree, and "coolant leakage" is at the 2nd level, then the level range involved is from the 2nd level to the 3rd level. This level range helps maintenance personnel quickly locate the approximate area where the fault hidden danger is located, and concentrate on in-depth analysis and troubleshooting of events at these levels, so as to more efficiently predict and prevent the occurrence of faults.

[0140] The beneficial effects of the above technology are that the leading occurrence probability determination unit determines the leading occurrence probability in combination with the occurrence probability of all intermediate events and top events and the decision leading comprehensive performance value of each level in the fault tree. This comprehensive consideration considers not only the possibility of the occurrence of the event itself, but also the leading role of the level in fault diagnosis, so that the evaluation of the event occurrence probability is more comprehensive and scientific, and can more accurately reflect the actual importance of the event in the entire fault system. The hidden danger prediction and evaluation unit takes the level range involved in the intermediate events and / or top events whose leading occurrence probability exceeds the threshold as the relative hidden danger level, provides maintenance personnel with an explicit and targeted fault hidden danger area, so that they can quickly focus on the key level where the fault hidden danger may exist, avoid blind troubleshooting, greatly improve the efficiency and accuracy of fault hidden danger prediction, help to develop a reasonable maintenance plan in advance, reduce the possibility of automobile fault occurrence, and ensure the safe and stable operation of the automobile.

[0141] Embodiment 9:

[0142] On the basis of embodiment 1, the maintenance strategy generation module refers to Figure 5 , and includes:

[0143] The maintenance item list retrieval sub-module is used to determine the highest level and the lowest level in the relative hidden danger level and the total number of covered levels, retrieve the maintenance item detail list corresponding to the fault phenomenon to be analyzed based on the highest level and the lowest level in the relative hidden danger level and the total number of covered levels, and determine all maintenance items and corresponding maintenance plan details;

[0144] The maintenance strategy generation sub-module is used to generate a maintenance strategy based on all maintenance items and corresponding maintenance plan details.

[0145] In this embodiment, the total number of covered levels refers to the sum of the number of levels involved in the intermediate events and / or top events in the fault tree of the fault phenomenon to be analyzed, whose leading occurrence probability exceeds the threshold value. For example, if the leading occurrence probability of events in the 2nd, 3rd and 5th levels in the fault tree exceeds the threshold value, the total number of covered levels is 3, which reflects the size of the level range spanned by the fault hidden danger in the fault tree and is an important indicator for evaluating the complexity of the fault and determining the maintenance scope.

[0146] In this embodiment, the maintenance item detail list of the fault phenomenon to be analyzed is a pre-arranged list that records in detail the highest level and the lowest level in different relative hidden danger levels corresponding to various possible fault phenomena to be analyzed, the total number of covered levels, the specific items that need to be maintained and the detailed information of each item. For example, when the highest level in the current relative hidden danger level of the automobile engine fault is 2, the lowest level is 5, and the total number of covered levels is 3, the maintenance item detail list indicates that the corresponding maintenance items include checking the spark plug, replacing the engine oil, detecting the valve clearance, and other maintenance items, as well as the operation steps, required tools, maintenance period, and other detailed contents of each item, providing a comprehensive reference basis for subsequent determination of specific maintenance work.

[0147] In this embodiment, all the maintenance items and the corresponding maintenance plan details are selected from the maintenance item detail list according to the total number of covered levels and the relative hidden danger level. The maintenance items clearly indicate the maintenance tasks that need to be performed for the current fault hidden danger, and the maintenance plan details further determine the specific execution scheme for each maintenance item, including maintenance time, maintenance personnel arrangement, maintenance resources required, etc. For example, it is determined that the maintenance items are replacing the brake pads and checking the brake oil, and the corresponding maintenance plan details may be that Zhang San, a maintenance technician, is responsible on Wednesday of this week, using the brake oil detection equipment and brake pad replacement tools in the repair workshop for operation, and preparing the corresponding type of brake pads and brake oil.

[0148] In this embodiment, generating the maintenance strategy based on all the maintenance items and the corresponding maintenance plan details is to integrate and plan the selected maintenance items and their maintenance plan details to form a complete and operable maintenance action scheme. This maintenance strategy takes into account the sequence of maintenance work, resource allocation, time arrangement, etc., to ensure that the maintenance work can be carried out efficiently and orderly. For example, according to the urgency and mutual dependence of the maintenance items, it is determined to first check the overall condition of the brake system, then replace the brake pads, and finally detect the brake oil; at the same time, human and material resources are reasonably allocated to ensure that each maintenance link can be smoothly implemented, thereby effectively preventing faults and ensuring the normal operation of the automobile.

[0149] The beneficial effects of the above technology are that the maintenance project list retrieval submodule retrieves the maintenance project detail list by determining the highest, lowest level and total number of coverage levels of the relative hazard level, accurately positioning all maintenance projects and corresponding maintenance plan details. This retrieval method based on the hazard level can find relevant maintenance projects according to the specific range of fault hazards, avoid comprehensive investigation without direction, and save manpower and time cost. The maintenance strategy generation submodule generates a maintenance strategy based on the determined maintenance projects and maintenance plan details, ensuring that the maintenance strategy is highly targeted and operable. Maintenance personnel can carry out maintenance work according to this strategy, improve maintenance efficiency, reduce vehicle downtime due to faults, extend the service life of the vehicle, ensure the safety and reliability of the vehicle operation, and improve the scientificity and systematicness of the overall maintenance management.

[0150] Embodiment 10:

[0151] The application provides an automobile fault diagnosis and predictive maintenance method based on a fault tree analysis method, comprising:

[0152] S1: constructing a fault tree of a fault phenomenon to be analyzed based on a top event of the fault phenomenon to be analyzed and all intermediate events and all basic events;

[0153] S2: determining all minimal necessary cut sets of the top event based on the fault tree of the fault phenomenon to be analyzed;

[0154] S3: calculating occurrence probability time sequence data of all intermediate events and the top event based on the loss trend data of the automobile components, and analyzing the relative hazard level of the fault tree of the fault phenomenon to be analyzed in a prediction period in combination with all minimal necessary cut sets of the top event;

[0155] S4: generating a maintenance strategy based on the relative hazard level of the fault tree of the fault phenomenon to be analyzed in the prediction period.

[0156] Obviously, those skilled in the art can make various modifications and variations to the application without departing from the spirit and scope of the application. Thus, if these modifications and variations of the application belong to the scope of the application and its equivalent technology, the application also intends to include these modifications and variations.

Claims

1. The automobile fault diagnosis and predictive maintenance system based on the fault tree analysis method is characterized by: include: A fault tree construction module is used to construct a fault tree of the fault phenomenon to be analyzed based on the top event, all intermediate events and all basic events of the fault phenomenon to be analyzed; A minimum cut set determination module is used to determine all minimum necessary cut sets of top events based on the fault tree of the fault phenomenon to be analyzed; The hidden danger prediction and assessment module is used to calculate the occurrence probability time series data of all intermediate events and top events based on the loss trend data of automobile parts, and combine all the minimum necessary cut sets of the top events to analyze the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period; The maintenance strategy generation module is used to generate a maintenance strategy based on the relative hidden danger level of the prediction period of the fault tree of the fault phenomenon to be analyzed.

2. The automobile fault diagnosis and predictive maintenance system based on the fault tree analysis method according to claim 1 is characterized in that: Fault tree building modules, including: An event determination submodule, configured to determine a top event, all intermediate events, and all basic events based on vehicle operation-related data of the fault phenomenon to be analyzed; The fault tree construction submodule is used to construct a tree structure for the top event, all intermediate events and all basic events of the fault tree of the fault phenomenon to be analyzed based on the logical relationship between all events, so as to obtain the fault tree of the fault phenomenon to be analyzed.

3. The automobile fault diagnosis and predictive maintenance system based on fault tree analysis method according to claim 1 is characterized in that: Minimum cut set determination module, including: The top event expression construction submodule is used to represent all basic events in the fault tree with Boolean variables, and to perform logical deduction step by step upward in the fault tree to obtain the Boolean expression of the top event; The expression simplification and expansion submodule is used to expand the Boolean expression of the top event based on the preset simplification rules to obtain multiple basic event product terms, and each basic event product term is regarded as a single minimum necessary cut set of the top event.

4. The automobile fault diagnosis and predictive maintenance system based on fault tree analysis method according to claim 1 is characterized in that: Hidden danger prediction and assessment module, including: The failure probability assessment submodule is used to determine the occurrence probability time series data of all basic events in the prediction period based on the loss trend data of automobile components; The event probability derivation submodule is used to derive the occurrence probability time series data of all basic events in the prediction period upward based on the fault tree to obtain the occurrence probability time series data of all intermediate events and top events; The fault synchronization value evaluation submodule is used to analyze the fault synchronization values ​​of all minimum necessary cut sets of top events based on the occurrence probability time series data of all basic events; The guidance performance evaluation submodule is used to assign a value to each level in the fault tree based on the fault synchronization value of all minimum necessary cut sets of the top event, and obtain the decision guidance comprehensive performance value of each level in the fault tree; The hidden danger prediction and evaluation submodule is used to determine the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period based on the occurrence probability of all intermediate events and top events and the decision-making guidance comprehensive performance value of each level in the fault tree.

5. The automobile fault diagnosis and predictive maintenance system based on the fault tree analysis method according to claim 4 is characterized in that: Failure probability assessment submodule, including: a first failure probability assessment unit, configured to perform segmented modeling based on a loss threshold for each automotive component and loss trend data of the automotive component, obtain a segmented failure rate function for each automotive component, and determine first failure probability data for each automotive component in a prediction period based on the segmented failure rate function for each automotive component; A second failure probability evaluation unit is configured to determine second failure probability data of each automobile component in a prediction period based on the future loss characteristics of each automobile component and a correlation model between the failure rate and the loss characteristics; a failure probability synthesis unit, configured to perform time-series alignment and weighted summation of first failure probability data and second failure probability data of each automobile component in a prediction period, to obtain final failure probability data of each automobile component in the prediction period; The event probability determination unit is used to determine the occurrence probability time series data of all basic events in the prediction period based on the final failure probability data of each automobile component in the prediction period.

6. The automobile fault diagnosis and predictive maintenance system based on fault tree analysis method according to claim 4 is characterized in that: The fault synchronization value evaluation submodule includes: A first synchronization value setting unit is configured to set the fault synchronization value of the corresponding minimum necessary cut set to 1 when a single minimum necessary cut set of a top event contains only one basic event; A same-order matrix building unit is used to synchronously divide and shuffle the occurrence probability time series data of all basic events in the corresponding minimum necessary cut set when a single minimum necessary cut set of the top event contains more than one basic event, obtain all occurrence probability sets of each basic event, and build the occurrence probability same-order matrix of each basic event in the corresponding minimum necessary cut set based on all occurrence probability sets of all basic events in the corresponding minimum necessary cut set; an eigenvector generating unit, configured to perform eigenvalue decomposition on each occurrence probability co-order matrix to obtain all eigenvalues ​​of each occurrence probability co-order matrix, and sort all eigenvalues ​​of each occurrence probability co-order matrix from largest to smallest to obtain an eigenvalue vector of each occurrence probability co-order matrix; The second synchronization value setting unit is used to calculate the similarity of all the same-row difference vectors, all the same-column difference vectors, and the eigenvalue vectors of the occurrence probability same-order matrix of all basic events in a single minimum necessary cut set, obtain multiple similarities of all basic events in the corresponding minimum necessary cut set, and determine the fault synchronization value of the corresponding minimum necessary cut set based on all the similarities of all basic events in the single small necessary cut set.

7. The automobile fault diagnosis and predictive maintenance system based on fault tree analysis method according to claim 4 is characterized in that: Boot performance evaluation submodule, including: an iterative synchronization value determination unit, configured to determine all common upper-level nodes of each minimum necessary cut set in the fault tree, and determine an iterative synchronization value of each minimum necessary cut set corresponding to each common upper-level node based on the level difference between the levels of all basic events included in each minimum necessary cut set in the fault tree and the corresponding levels of each common upper-level node and the fault synchronization value of the corresponding minimum necessary cut set; a relative floor height ratio determination unit, configured to determine a relative floor height ratio of each common upper-layer node of each minimum necessary cut set among all common upper-layer nodes of the corresponding minimum necessary cut set; The decision guidance comprehensive performance determination unit is used to determine the decision guidance comprehensive performance value of each level in the fault tree based on the iterative synchronization value of all common upper-level nodes contained in each level in the fault tree and the relative layer height ratio of all common upper-level nodes in the corresponding minimum necessary cut set.

8. The automobile fault diagnosis and predictive maintenance system based on fault tree analysis method according to claim 4 is characterized in that: Hidden danger prediction and assessment submodule includes: A guiding occurrence probability determination unit is used to determine the guiding occurrence probability of all intermediate events and the top event based on the occurrence probability of all intermediate events and the top event and the decision guidance comprehensive performance value of each level in the fault tree; The hidden danger prediction and evaluation unit is used to regard the hierarchical range involved in the corresponding fault tree of the fault phenomenon to be analyzed, including all intermediate events and / or top events whose guiding probability of occurrence exceeds a threshold, as the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period.

9. The automobile fault diagnosis and predictive maintenance system based on fault tree analysis method according to claim 1, characterized in that: Maintenance strategy generation module, including: The maintenance item list retrieval submodule is used to determine the highest and lowest levels in the relative hidden danger hierarchy and the total number of covered levels. Based on the highest and lowest levels in the relative hidden danger hierarchy and the total number of covered levels, it retrieves a detailed list of maintenance items corresponding to the fault phenomenon to be analyzed, and determines all items that require maintenance and the corresponding maintenance plan details; The maintenance strategy generation submodule is used to generate a maintenance strategy based on all maintenance items and corresponding maintenance plan details.

10. The automobile fault diagnosis and predictive maintenance method based on the fault tree analysis method is characterized by: include: S1: Construct a fault tree of the fault phenomenon to be analyzed based on the top event, all intermediate events, and all basic events of the fault phenomenon to be analyzed; S2: Determine all minimum necessary cut sets of top events based on the fault tree of the fault phenomenon to be analyzed; S3: Based on the wear trend data of automotive components, the occurrence probability time series data of all intermediate events and top events are calculated. By combining all the minimum necessary cut sets of the top events, the relative hidden danger level of the fault tree of the fault phenomenon to be analyzed in the prediction period is analyzed. S4: Generate maintenance strategies based on the relative hidden danger levels of the fault tree of the fault phenomenon to be analyzed during the prediction period.

Images